US7853356B2 - Method for optimizing a robot program and a robot system - Google Patents

Abstract

An apparatus and a method for optimizing robot performance includes a computer connected to the robot controller for receiving performance data of the robot as the controller executes a path program. The computer uses the performance data, user specified optimization objectives and constraints and a kinematic/dynamic simulator to generate a new set of control system parameters to replace the default set in the controller. The computer repeats the process until the new set of control system parameters is optimized.

Description

BACKGROUND OF THE INVENTION

The present invention relates generally to a method and an apparatus for modifying a robot control program to meet performance objectives.

The closest prior art to this invention is based on the following principle:

A robot controller connects to an external computing device such as a personal computer through communication link. The external device (e.g. a personal computer or “PC”) accesses the memory area of robot controller. User programs stored in the memory area can therefore be manipulated by the external device.

The main concept of this prior art device is for either connecting multiple robots to transfer user programs or for data storage purposes. There is no real-time interaction between the robot controller and the external computing device for path performance optimization.

SUMMARY OF THE INVENTION

The present invention concerns a method and an apparatus for robot controller operation optimization. The present invention does not only use a communication channel to connect the robot controller and the external computing device (a general PC), but makes use of the CPU power of the external PC to analyze and optimize the robot path real-time. This external PC becomes a highly flexible, re-configurable, and yet powerful second processor for the robot controller.

The method and apparatus according to the present invention are extremely useful for small shape cutting/generation because the quality of these types of processes is tedious and time-consuming to examine and verify. Also, the method and apparatus can optimize robot path and cycle time for load-unload and other material handling applications successfully. Furthermore, the method and apparatus lay the perfect groundwork for extended robot optimization in other applications such as palletizing, spot welding, etc.

DESCRIPTION OF THE DRAWINGS

The above, as well as other advantages of the present invention will become readily apparent to those skilled in the art from the following detailed description of a preferred embodiment when considered in the light of the accompanying drawings in which:

FIG. 1 is a block diagram of a robot program optimization apparatus in accordance with the present invention;

FIG. 2 is block diagram of the robot controller shown inFIG. 1;

FIG. 3 is a block diagram of the external personal computer shown inFIG. 1; and

FIG. 4 is a flow diagram of the method according to the present invention.

DESCRIPTION OF THE PREFERRED EMBODIMENT

Traditionally, a robot program is developed and occasionally optimized independently of the robot system to meet some performance objectives such as higher accuracy or better cycle time. When the program is executed on different robots, that are not usually exactly identical, the robots' performances usually vary and some may not meet the desired objectives. This variation is attributed to the need of different manipulators for different “system parameters” such as acceleration time, over-current protection limits, servo loop gains, coulomb friction parameters, integration gains, spring constants, etc, that are usually fixed for specific manipulator models.

The present invention recognizes that differences between manipulators and operating conditions affect robot performance and provides a method to optimize the robot program as well as the system parameters according to the robot system (control and manipulator) and the operating conditions. The robot program, which may be developed in traditional ways, is executed by the robot system for which certain parameters, usually encoder positional data and motor current data, are monitored and communicated during robot operation to a secondary processor for optimization. During the operation of the robot, the secondary processor applies optimization routines to meet predetermined objectives, and modifies not only the robot program but also the control system parameters. The optimization takes into consideration the overall robot model, not just individual drive axes, and monitors the trend of specific composites of performance parameters to arrive at optimum control system parameters. This process is repeated while the robot is operational until the optimization objectives are met at which time the program is finalized and the system parameters are set for productive operation.

The related prior art includes offline programming in which the robot program is modified using a secondary processor, usually offline, to meet certain objectives. The modifications are done independently of the control system parameters and are not checked in real-time against actual performance. Accordingly, the same program may be executed differently by different manipulators or by the same manipulator under different operating conditions.

Learning control also modifies the robot program in real-time and may modify some control system parameters. However, in learning control optimizing iterations are run on an interval by interval basis where the state condition of the servo system changes with risk of instability. With its inherent slow conversion process only a few parameters can be iterated, usually servo gains and damping coefficients.

The present invention iterates at a fixed state on the PC for faster conversion and no risk of instability. This also allows more control parameters to be optimized within a given time period, hence better optimization of performance.

Unlike the present invention, none of the prior art methods accommodate input from the motor current that is essential to avoid overloading the motors and degrading the durability of the robot. Neither do they accommodate changes in the spring constants hence compromising vibration performance of the robot.

The present invention uses an iterative approach to evaluate predictive changes to the control parameters on the PC, using the dynamic robot model. Unlike learning control which must use the actual robot to evaluate the effect of changes, the present invention can run the model on the PC many more times under the same conditions and only applies the change once an objective is attained. This fast iterative approach also avoids the risk of instability.

Accordingly, the present invention combines the advantages of offline programming, where the program modifications can be run independently of the robot, with learning control where iterative operations lead to optimized parameters. In learning control the iteration is run on an interval by interval basis where the state condition of the servo system changes with risk of instability, but the present invention iterates at a fixed state for faster conversion and no risk of instability. Furthermore, in learning control, a change in one parameter can not be evaluated under the same dynamic state condition until the whole program has been executed, a very slow process. Also, the approach of learning control is compromised by making the evaluation on the following time interval under a new state condition. This makes stability of the iteration difficult to attain.

Robot path optimization is a CPU-intensive task and is highly dependent on the robot TCP position and robot arm configuration. In addition, the CPU in a robot controller usually handles excessive tasks such as motion planning, program management, and memory management etc. Therefore, it has been difficult to achieve path optimization with the robot controller's main CPU in the past. The method and apparatus according to the present invention provide the feasible measure to make real-time robot path optimization possible.

The main concept of this invention is to interact an external computing device (a general PC) with a robot controller real-time. The feedback of motion performance can be displayed to users on any PC with a network card through either a dedicated connection between the robot controller and the PC, or via a network. This visual display can provide useful information from the controller such as path deviation and cycle time with or without the actual process taking place.

To visualize motion performance and display to the user every time when the program executes, critical motion data and system/servo status will be transmitted to the PC real-time. Also, each record will be analyzed and stored for future reference based on the user's desire. After the analysis, the CPU power of the external PC will be used to calculate the compensation data for the optimization iteration. This important compensation data will then be transmitted back to the robot controller real-time and the next iteration trial will be started automatically. This process will continue until the user-defined criterion is met.

There is shown inFIG. 1 anapparatus10 for the optimization of a robot control program in accordance with the present invention. A first control means11, such as a robot controller, is connected to a second control means12, such as a personal computer (PC), by acommunication link13 such as a computer network. Although a PC is preferred for the second control means, any suitable computer can be used. Thecontroller11 includes amotion system14, aservo system15 and a communication client/server16. The PC12 includes apath analysis module17, anoptimization module18 and a communication client/server19.

As shown inFIG. 2, in therobot controller11 themotion system14 is connected to theservo system15 and executes a control program to generate “motion commands” and “adjust servo system” signals to theservo system15. Theservo system15 is connected tomotors20 of the robot and executes the motion commands to operate themotors20 and receives feedback signals including motor current from themotors20. Multiple control programs can be stored in auser programs memory21 connected to the communications server/client16. Themotion system14 sends “performance data” to and receives “optimized data” from the communication server/client16. The “optimized data” is used to generate the “adjust servo system” signals to optimize performance.

Theexternal PC12 is shown inFIG. 3 as including the communication server/client19 that is connected to deliver the “performance data” to aperformance data memory22. A kinematic/dynamic models memory23 stores such models of the robot. Thedata memory22 and themodels memory23 are connected to arobot performance module24 to “establish” the actual performance of the robot. Themodule24 is connected to an analyze path withsimulator module25 that utilizes simulation to analyze the robot path based upon the robot performance. The result from themodule25 is provided to anoptimization iteration module26. The “optimized data” from themodule26 is returned to thesimulator module25 via a feedback loop to check the path based upon the “optimized data”. Upon confirmation that the “optimized data” produces better performance of the robot, the “optimized data” is sent in real time to the motion system ofFIG. 2 through the communication server/clients19 and16 of the communications link13.

FIG. 4 is a flow diagram for the method according to the present invention. The method begins at a “Connect to Robot Controller”instruction set30 whereby theexternal PC12 connects to therobot controller11 as shown inFIG. 1. Then the PC executes a “Synchronize with controller”instruction set31 wherein thePC12 is synchronized in real time with the operation of therobot controller11. ThePC12 receives data from thecontroller11 in a “Receive performance data from controller”instruction set32. The execution of theinstruction set31 also causes execution of a “Create kinematic/dynamic simulator”instruction set33. The performance data received in thestep32 and the simulator created in thestep33 are used to execute an “Analyze robot performance”instruction set34. Thestep34 leads to a “User specifies optimization objectives and constraints”instruction set35. Thesteps33 and35 lead to an “Optimize using user program and simulator”instruction set36 that generates a potential optimized program. At “Satisfy objectives?”decision point37, if the optimized program does not satisfy the user specified objectives, the method branches at “N” to return to theoptimizer step36. If the optimized program does satisfy the user specified objectives, the method branches at “Y” to a “Transmit necessary system parameters and/or user programs back to controller”instruction set38 and the optimization process is complete. Now therobot controller11 can execute the optimized program and/or change the control system parameters.

In accordance with the provisions of the patent statutes, the present invention has been described in what is considered to represent its preferred embodiment. However, it should be noted that the invention can be practiced otherwise than as specifically illustrated and described without departing from its spirit or scope.

Claims (16)

1. A robot control system, comprising:

a robot;

a first control means including a controller connected to said robot and executing a control program, said robot being operated in accordance with control system parameters specified in said control program;

means for monitoring actual performance data of said robot during the operation of said robot; and

a second control means connected to said controller and to said means for monitoring, said second control means being responsive to said actual performance data for automatically applying optimization routines to a kinematic/dynamic simulator to modify said control system parameters in said control program in real-time while said robot is operated by said controller.

2. The robot control system according toclaim 1 wherein said second control means is a personal computer.

3. The robot control system according toclaim 1 wherein said controller, said means for monitoring and said second control means are connected by a communication link for data transfer.

4. The robot control system according toclaim 3 wherein said communication link is a computer network.

5. The robot control system according toclaim 1 wherein said means for monitoring includes means for generating feedback signals from motors operating said robot.

6. The robot control system according toclaim 5 wherein said feedback signals include motor currents of said motors.

7. The robot control system according toclaim 1 wherein said second control means includes user specified optimization objectives and constraints for modifying said control system parameters.

8. In a robot system including a manipulator and a controller, a method for optimizing a robot path program within given objectives and accommodating variations in dynamic performance among otherwise similar robots, comprising the steps of:

a. generating a path program for a robot by a conventional robot teaching method;

b. providing a computer with a dynamic model of the robot and its control system parameters and an optimization program;

c. providing the computer with target performance objectives;

d. connecting the controller for communication with the computer;

e. operating the robot by the controller executing the path program under a default set of control system parameters;

f. monitoring a set of performance variables of the robot during operation;

g. communicating the performance variables to the computer;

h. executing the optimization program in the computer using the dynamic model and the performance variables to meet the target performance objectives;

i. generating a new set of control system parameters;

j. communicating the new set of control system parameters to the controller;

k. assigning the new set of control system parameters to the path program as the default set of control system parameters;

l. repeating said steps e. through k. until the new set of control system parameters is substantially similar to the default set of control system parameters resulting in an optimized default set of control system parameters, wherein said control system parameters are automatically modified in said control program in real-time while said robot is operated by said controller; and

m. disconnecting the computer from the controller as the controller continues to operate the robot by executing the path program under the optimized default set of control system parameters.

9. The method accordingclaim 8 wherein one of the target performance objectives is a cycle time of the robot.

10. The method according toclaim 9 wherein the performance variables include axis motor current and axis encoder counts.

11. The method accordingclaim 8 wherein one of the target performance objectives is robot path accuracy.

12. The method according toclaim 11 wherein the performance variables include axis encoder errors, axis motor current, axis encoder counts, and servo gains.

13. A robot control system, comprising:

a robot;

a robot controller connected to said robot and executing a control program, said robot being operated in accordance with control system parameters specified in said control program;

means for monitoring actual performance data of said robot during the operation of said robot; and

a computer connected to said robot controller and to said means for monitoring by a communication link, said computer being responsive to said actual performance data for automatically applying optimization routines to a mathematical simulation model to modify said control system parameters in said control program in real-time while said robot is operated by said controller.

14. The robot system according toclaim 13, further comprising an optimization program embedded into said second control means for analyzing said actual performance data and interacting with said mathematical model and said robot to test and modify said control system parameters until desired performance objectives are realized by said robot.

15. The robot control system according toclaim 13, wherein said means for monitoring includes the mathematical simulation model of said robot.

16. The robot system according toclaim 15, further comprising an optimization program embedded into said computer for analyzing said actual performance data and interacting with said mathematical model and said robot to test and modify said control system parameters until desired performance objectives are realized by said robot.

Images